Battery Discharge Results Using Modified Andriod App

[Liger 1956]

I have done another test with my 2 batteries. Battery 1 came with the Anafi and I bought Battery lightly used from another forum member a while ago. I did a battery reset on Battery 2 to see whether the figures would alter. Neither battery has been heavily used or run down to RTF or autoland voltages.

The batteries show a good corrolation in both % and voltage. It was a still day today and about 8 degrees. The Battery 2 test pre reset was done in about slight wind at the same temperature. The flights were made within 1 hour of the batteries being charged throgh the USB-C port using a PD wall charger. All of the tests are a simple hover 1m above ground for 15 minutes and then land. Both battery voltages have recovered to about 7.6v.

I am beginning to think that the % battery reported by FF6 6.6.2 is related more to the time of flight than the actual battery voltage. If we assume that the battery will go from a reported 100% to 0% in 25minutes (Parrot advertised maxiumum flight time) then this equates to 4% used/min which is approximately what the graphs above are showing.

As an aside I did have 2 cases, both with battery 2, when the gimbal would not initialise properly and it ended up facing the sky. It has done this once before a long while ago. Rebooting the Anafi cured the problem.

 

[Jagerbomb52]

Just to add to your battery initialize problem I also have noticed a few times my camera will not initialize and I have to slide the battery in and out a few times to get the camera to initialize.

 

[Gipsz Jakab]

I have done another test with my 2 batteries. Battery 1 came with the Anafi and I bought Battery lightly used from another forum member a while ago. I did a battery reset on Battery 2 to see whether the figures would alter. Neither battery has been heavily used or run down to RTF or autoland voltages.
View attachment 5127
The batteries show a good corrolation in both % and voltage. It was a still day today and about 8 degrees. The Battery 2 test pre reset was done in about slight wind at the same temperature. The flights were made within 1 hour of the batteries being charged throgh the USB-C port using a PD wall charger. All of the tests are a simple hover 1m above ground for 15 minutes and then land. Both battery voltages have recovered to about 7.6v.

I am beginning to think that the % battery reported by FF6 6.6.2 is related more to the time of flight than the actual battery voltage. If we assume that the battery will go from a reported 100% to 0% in 25minutes (Parrot advertised maxiumum flight time) then this equates to 4% used/min which is approximately what the graphs above are showing.

As an aside I did have 2 cases, both with battery 2, when the gimbal would not initialise properly and it ended up facing the sky. It has done this once before a long while ago. Rebooting the Anafi cured the problem.

"I am beginning to think that the % battery reported by FF6 6.6.2 is related more to the time of flight than the actual battery voltage"
I reached the same opinion earlier:
"The lesson of the table: the% display of the battery charge does not seem to reflect the change in voltage, it is too linear. I think it has more to do with the elapsed time than the battery voltage."

 

[Gipsz Jakab]

Just to add to your battery initialize problem I also have noticed a few times my camera will not initialize and I have to slide the battery in and out a few times to get the camera to initialize.

The same problem often occurs. The initialization does not run properly, then starts turning upwards to "protest" violently, shaking the camera left and right.
Turning it off and on again solves it ...
This I think has been released since the 1.6.2 update, somewhere someone wrote this.

 

[BRC]

I have done another test with my 2 batteries. Battery 1 came with the Anafi and I bought Battery lightly used from another forum member a while ago. I did a battery reset on Battery 2 to see whether the figures would alter. Neither battery has been heavily used or run down to RTF or autoland voltages.
View attachment 5127
The batteries show a good corrolation in both % and voltage. It was a still day today and about 8 degrees. The Battery 2 test pre reset was done in about slight wind at the same temperature. The flights were made within 1 hour of the batteries being charged throgh the USB-C port using a PD wall charger. All of the tests are a simple hover 1m above ground for 15 minutes and then land. Both battery voltages have recovered to about 7.6v.

I am beginning to think that the % battery reported by FF6 6.6.2 is related more to the time of flight than the actual battery voltage. If we assume that the battery will go from a reported 100% to 0% in 25minutes (Parrot advertised maxiumum flight time) then this equates to 4% used/min which is approximately what the graphs above are showing.

As an aside I did have 2 cases, both with battery 2, when the gimbal would not initialise properly and it ended up facing the sky. It has done this once before a long while ago. Rebooting the Anafi cured the problem.

What might be interesting to see is results under actual flight conditions - maybe with a flight plan so it could be repeated with different batteries
Speed, wind conditions, loads, environment, etc. must have an impact on battery usage & current and should show a nonlinear drop of percentage of battery usage, along with voltage
Is this not correct?

 

[Gipsz Jakab]

What might be interesting to see is results under actual flight conditions - maybe with a flight plan so it could be repeated with different batteries
Speed, wind conditions, loads, environment, etc. must have an impact on battery usage & current and should show a nonlinear drop of percentage of battery usage, along with voltage
Is this not correct?

I don’t think there will be much difference because hovering in one place is energy intensive due to a phenomenon called tranlation lift caused by turbulence. Only by flying against the wind could you constantly load it so that it is really visible on the graph. But how would you do that? If you start against the wind and turn around at about 60-70%, the load will be lower than usual with the backwind. If you load upwards, you will lose weight when you come down. I think you can make the voltage curve wavy, but by the end it’s about the same result as floating.
I didn’t try all of this, but with the help of logic, I played it in my head. It will be interesting to see because we will definitely try it sometime ...

 

[MustangPhil]

Bud made an interesting video with 2 flights And 3 Auto-landings with the modified FF6 where you can see the voltage displayed in real time.

The purpose of his flights was to test different 18650 batteries and his specific setting...

But to connect it to the actual thread, I remarked something very common to the 3 auto-landings :
All of auto-landings are triggered when the voltage reaches a value under 6600 mV whatever the % displayed.
This is something that may also be visible on the curves of this thread...

For info, his test is here :

 

[Liger 1956]

I have a theory that the auto RTF is triggered by an algorithm based on the remaining battery % and the distance of the Anafi from the home point whilst the auto land is triggered by remaining voltage. If that is correct it could be the reason why people have had their Anafi autoland before the RTF was triggered especially if the battery used had a low battery health.

 

[Gipsz Jakab]

I have a theory that the auto RTF is triggered by an algorithm based on the remaining battery % and the distance of the Anafi from the home point whilst the auto land is triggered by remaining voltage. If that is correct it could be the reason why people have had their Anafi autoland before the RTF was triggered especially if the battery used had a low battery health.

I agree with you.
However, this flight also showed that the battery level% display has nothing to do with the voltage, it starts at 84% and lands 45-55% ... that's not right ...
The correct values are 7629mV = 45% and 6500mV = 1%
Real flight 85% -46%
Theoretical correct 45% -1%
A total of approx. 30-35% flight time compared to LiHV cell

 

[Bud FPV]

I have a theory that the auto RTF is triggered by an algorithm based on the remaining battery % and the distance of the Anafi from the home point whilst the auto land is triggered by remaining voltage. If that is correct it could be the reason why people have had their Anafi autoland before the RTF was triggered especially if the battery used had a low battery health.

Personally I guess the %battery come from the current drain (the comsumation with current sensor). Need to see how ardupilot calcul that .

 

[Bud FPV]

The Anafi come from ardupilot but modified Battery Failsafe — Copter documentation

 

[sibujacob]

I don’t think there will be much difference because hovering in one place is energy intensive due to a phenomenon called tranlation lift caused by turbulence. Only by flying against the wind could you constantly load it so that it is really visible on the graph. But how would you do that? If you start against the wind and turn around at about 60-70%, the load will be lower than usual with the backwind. If you load upwards, you will lose weight when you come down. I think you can make the voltage curve wavy, but by the end it’s about the same result as floating.
I didn’t try all of this, but with the help of logic, I played it in my head. It will be interesting to see because we will definitely try it sometime ...

While reading these facts about deviations seen in battery % levels and actual battery voltage levels, the one thing i could recommend here to Parrot is make a AND logic condition in their BMS to check if flight time is established near to battery % or battery voltage and then take the decision for the RTH or auto land trigger.

 

[Ana789]

I don’t think there will be much difference because hovering in one place is energy intensive due to a phenomenon called tranlation lift caused by turbulence. Only by flying against the wind could you constantly load it so that it is really visible on the graph. But how would you do that? If you start against the wind and turn around at about 60-70%, the load will be lower than usual with the backwind. If you load upwards, you will lose weight when you come down. I think you can make the voltage curve wavy, but by the end it’s about the same result as floating.
I didn’t try all of this, but with the help of logic, I played it in my head. It will be interesting to see because we will definitely try it sometime ...

And in case you "play" it again, in your head, put "a windless day and a constant speed of 20ish kph" in there, too?
Every direction it creates the same "flow against the wind", so you can fly it in endless big circles till the battery starts the landing.
(or till it fells down ;-)

 

[Gipsz Jakab]

And in case you "play" it again, in your head, put "a windless day and a constant speed of 20ish kph" in there, too?
Every direction it creates the same "flow against the wind", so you can fly it in endless big circles till the battery starts the landing.
(or till it fells down ;-)

I don't really understand that answer. Could it be the language difficulty?

 

[Ana789]

I don't really understand that answer. Could it be the language difficulty?

Can, that is always a possibility!

I can try to help you over your difficulties:

On a windless day, with a constant speed, you have always the same translational lift, it does not matter the direction you fly.
And since this translational lift does just need movement of the bird, not additional wind from the front, you "just" need to "find" the optimal speed.
Which 20kph , or a bit more, seems to be it, for most drones.
Since that is often the mentioned condition for longest flight times, in theory at least!

 

[Gipsz Jakab]

Can, that is always a possibility!

I can try to help you over your difficulties:

On a windless day, with a constant speed, you have always the same translation lift, it does not matter the direction you fly.
And since this translation lift does just need movement of the bird, not additional wind from the front, you "just" need to "find" the optimal speed.
Which 20kph , or a bit more, seems to be it, for most drones.
Since that is often the mentioned condition for longest flight times, in theory at least!

I understand the words, but it doesn’t put together what you want to say to me. So I can't confirm or deny ...

 

[Ana789]

I understand the words, but it doesn’t put together what you want to say to me. So I can't confirm or deny ...

Dont you find that a bit strange?
You pointed to the translational lift, but can't put together, that it is available also on windless days and what that means for flight times?
Perhaps you look it up at wiki?

"Improved rotor efficiency resulting from directional flight in a helicopter is called translational lift. The efficiency of the hovering rotor system is greatly improved with each knot of incoming wind gained by horizontal movement of the aircraft or surface wind"

This gets us , in smooth big circles, the optimal powerusage in max flight time tests!

 

[Gipsz Jakab]

Dont you find that a bit strange?
You pointed to the translational lift, but can't put together, that it is available also on windless days and what that means for flight times?
Perhaps you look it up at wiki?

"Improved rotor efficiency resulting from directional flight in a helicopter is called translational lift. The efficiency of the hovering rotor system is greatly improved with each knot of incoming wind gained by horizontal movement of the aircraft or surface wind"

This gets us , in smooth big circles, the optimal powerusage in max flight time tests!

Of course, I understand all this, since I mentioned it first. I don't understand why we're talking about this now?
It was about batteries. The pilot performing the tests described that he took the tests in a floating position. To this, someone replied that he would be curious how the battery charge would change during the flight. To this I replied that hovering requires more energy than flight due to the phenomenon of translational lift, so no shorter battery time is expected during flight. If the headwind puts more strain on it at first, it will turn around and come home with a tailwind, so even then the load will be evened out.
I have already described all this once. So now what is that, I did not write well, or what is the question? Of course I’m not angry, I just want to understand what you want to talk about.

 

[Ana789]

My point is:

"Only by flying against the wind could you constantly load it so that it is really visible on the graph. But how would you do that? If you start against the wind and turn around at about 60-70%, the load will be lower than usual with the backwind. ...."

But how would you do it?
A flight, when there is NO WIND, is the only chance to get competitive results with all batteries, someone is using!
Using the speed, the manufacturer is giving out for the max. flight time, will give you the neededlift. And flying big circles (by hand on anafi, since a flight plan is kind of a "stop'n'go" flight at every waypoint!), and b/c I can't find any "max flight time at no wind and constant xx kph flight time" for Anafi, I'd say it is full speed in "film" mode with this bird.
Or worst, some speed higher as "max film", but lower as "max sport", so difficult to hit all the time.

I have not done that with the Anafi, but I did that with a DJI Spark, some years ago.
It was new, I wanted to fly the most distance and the longest flighttime I could get out of one battery, so I did that in a "no wind" day, on top of a small hill.
One battery for max flight time in normal mode, and one battery in Sport for max flight distance, doing big circles around the hill (at that time, litchi still wasn't available for the Spark).
I got 15+ minutes out from the first battery, and just about 9 km from the second one.
Tried the flight time thing on a windy day: Never got really 14+ minutes!

So the answer (imho) is: No wind, find the speed for the self created translational lift and check it out.

Merry X-mas!

 

[jaja6009]

Some more statistics to add.

Very short autonomous flight to document construction progress. Very windy and the Anafi had to "lean into" the wind. This is why the battery voltage is lower at 90 compared to 80 and 70%.

Battery Info: 18 Flights, 3.5 Flight Hours, 21 charges to 100%, Average % discharged to: 43%
On Freeflight 6 Battery Capacity: 88%

Power on Drone: 97% 8406 mV

Percentage​	Voltage​
90​	7532​
80​	7693​
70​	7549​

As stated, the aircraft was flying into 10 to 15 mph winds at 90%. At 80% it was flying lateral to the wind.

Something I am seeing is that my previous ignorance on battery care has my batteries in worse condition than their age, times charged and times used would indicate. Lesson learned.

And I repeat it again, the voltage Mod has saved my Anafis. I am very confident flying them again.